Water Cooling a Lost Art?

[bigben2k]

Well, now that you're on your way with a testbench, I think you'd be able to experiment with different designs, and show actual results.

You can expand on a design, by making numerous iterations of it, and make very small changes, to basically optimize a particular design.

I'd like to see a Cascade style block development; I'd love to do it myself, but I'm nowhere near having all the equipment. The Cascade style block is the most challenging one, if you ask me. I'd like to see various cup spacing, different inlet diameters and spacing. It's a lot of work, and can involve working with some very fine parts.

[Delirious]

Originally Posted by jaydee116

No where, that's the problem. Looking for ideas on what direction to go.... I have no spacific intrest in any one part of it. I just want to find one part of it that would be most useful for everyone here at ProCooling and stay on track with it. What part that is I am not sure.

What are people most interested in here?

Id really like to see some solid radiator testing and design. It seems like we have pushed the water blocks to marginal gains in each new design but the radiator/heatercore seems neglected. Id really like to see how a bunch of popular radiators/heatercores that everyone uses today stand up against each other.

Direct die cooling seems like it would produce excellent results, but would definetly be even harder to pull off than phase change.

[jaydee]

Originally Posted by bigben2k

Well, now that you're on your way with a testbench, I think you'd be able to experiment with different designs, and show actual results.

You can expand on a design, by making numerous iterations of it, and make very small changes, to basically optimize a particular design.

I'd like to see a Cascade style block development; I'd love to do it myself, but I'm nowhere near having all the equipment. The Cascade style block is the most challenging one, if you ask me. I'd like to see various cup spacing, different inlet diameters and spacing. It's a lot of work, and can involve working with some very fine parts.

Out of my capability unfortunately. Also who would use the information gathered? Not going to further assist manufactures with free R&D. That kind of stuff is way out of the DIY league anyway.

Originally Posted by Delirious

Id really like to see some solid radiator testing and design. It seems like we have pushed the water blocks to marginal gains in each new design but the radiator/heatercore seems neglected. Id really like to see how a bunch of popular radiators/heatercores that everyone uses today stand up against each other.

Direct die cooling seems like it would produce excellent results, but would definitely be even harder to pull off than phase change.

I am not touching Rads, maybe pH can do that. To complicated to test for me and no way to make a prototype on any design we came up with. The prototype would be commercial anyway. I know of 3 outfits that already have superior radiator designs anyway. The only hold up is lack of market. The cost of tooling up to make such a thing would be well more than money made from it.

I have already done direct die cooling. No reason to pursue it.

[Cathar]

Originally Posted by jaydee116

No where, that's the problem. Looking for ideas on what direction to go.... I have no spacific intrest in any one part of it. I just want to find one part of it that would be most useful for everyone here at ProCooling and stay on track with it. What part that is I am not sure.

What are people most interested in here?

Some ideas that have been at the fore-front of my mind, and they've all had some good progress:

1) Waterblock design (as always) - getting pretty damn close to a wall here at the upper end of the flow rate/pressure supply range (for pumps that aren't ridiculous in size/power), but there really are substantial gains to go for everyone to improve lower pressure/flow performance. This will go a LONG way towards facilitating even smaller (8mm or 5/16" ID) tubing use such that the performance trade-offs for using such are very minimal while allowing for easier routing in-case.

2) Radiator design - The Thermochill PA160 is the culmination of a lot of ideas already discussed here at Procooling and that's about to hit the scene. A joint effort radiator design that marries low-noise fan use, compactness and fairly easy in-case installability. Improvements from here on would be slowly incremental in nature. There's still some ways to go, and lots of testing, and if everyone helped out here in some way it'd be great.

3) Pump advances - we've all seen some really excellent advances here. The latest actually quiet Laing D4's with the speed adjustable knobs on the back that accept from 8v-24v input, and at 12v match the old D4's but without all the noise are a tremendous leap in the right direction. The Iwaki RD-30 quite possibly still remains the king of the heap, but with these latest D4's it's no longer a one-horse show. The Laing DDC also remains as an excellent choice - fix the inlet issue and it would quite possibly be the pump that "has it all" when size is taken into account.

4) Chilled water use - there are a number of exciting opportunities here. I have a very cheap to make prototype here that could fit in a CD-ROM bay, and am co-ordinating with an electronics whizz-kid to provide for dew-point temperature management while being powered from a regular PC PSU. There are other approaches and other designs out there on the market as well, but there's lot of work to do to produce something that is capable of chilling water with a good COP in a compact area. Targetting such a goal drives water-cooling's viability as keeping far ahead of anything that air/heat-pipe cooling can achieve.

In a computing world whereby top-end CPU's are increasingly bin-split according to ability to run stably at higher speeds, and charged accordingly, the ability to sneak some solution in there which enables a cheaper low-end CPU to easily surpass the speed of a higher end CPU for less noise than the stock cooling solution on that higher-end CPU produces, and for less cost keeps that side of the market alive.

With multi-cored CPU's and CPU manufacturers increasingly focusing on thermal-management within the CPU dies to run CPU's part-time at their full speed when possible, a thermal cooling solution that enables full-time full-speed is still a worth-while goal within the community.

Where it gets dicey, as always, is with the ability to do it all cost-effectively and cost-justifiably for the mass-market. Sadly, this is something that DIY'ers will never be able to achieve.

[jaydee]

Originally Posted by Cathar

Some ideas that have been at the fore-front of my mind, and they've all had some good progress:

1) Waterblock design (as always) - getting pretty damn close to a wall here at the upper end of the flow rate/pressure supply range (for pumps that aren't ridiculous in size/power), but there really are substantial gains to go for everyone to improve lower pressure/flow performance. This will go a LONG way towards facilitating even smaller (8mm or 5/16" ID) tubing use such that the performance trade-offs for using such are very minimal while allowing for easier routing in-case.

2) Radiator design - The Thermochill PA160 is the culmination of a lot of ideas already discussed here at Procooling and that's about to hit the scene. A joint effort radiator design that marries low-noise fan use, compactness and fairly easy in-case installability. Improvements from here on would be slowly incremental in nature. There's still some ways to go, and lots of testing, and if everyone helped out here in some way it'd be great.

3) Pump advances - we've all seen some really excellent advances here. The latest actually quiet Laing D4's with the speed adjustable knobs on the back that accept from 8v-24v input, and at 12v match the old D4's but without all the noise are a tremendous leap in the right direction. The Iwaki RD-30 quite possibly still remains the king of the heap, but with these latest D4's it's no longer a one-horse show. The Laing DDC also remains as an excellent choice - fix the inlet issue and it would quite possibly be the pump that "has it all" when size is taken into account.

4) Chilled water use - there are a number of exciting opportunities here. I have a very cheap to make prototype here that could fit in a CD-ROM bay, and am co-ordinating with an electronics whizz-kid to provide for dew-point temperature management while being powered from a regular PC PSU. There are other approaches and other designs out there on the market as well, but there's lot of work to do to produce something that is capable of chilling water with a good COP in a compact area. Targetting such a goal drives water-cooling's viability as keeping far ahead of anything that air/heat-pipe cooling can achieve.

In a computing world whereby top-end CPU's are increasingly bin-split according to ability to run stably at higher speeds, and charged accordingly, the ability to sneak some solution in there which enables a cheaper low-end CPU to easily surpass the speed of a higher end CPU for less noise than the stock cooling solution on that higher-end CPU produces, and for less cost keeps that side of the market alive.

With multi-cored CPU's and CPU manufacturers increasingly focusing on thermal-management within the CPU dies to run CPU's part-time at their full speed when possible, a thermal cooling solution that enables full-time full-speed is still a worth-while goal within the community.

Where it gets dicey, as always, is with the ability to do it all cost-effectively and cost-justifiably for the mass-market. Sadly, this is something that DIY'ers will never be able to achieve.

A lot of what you have on the pate here is already being developed by many companies. Main hold up is there isn't the market to warrant putting it into production.

IF water cooling goes main stream (as in every computer manufacturer's standard comps have water cooling) all this stuff will start flying out of the wood work like crazy. We will see stuff we would have not thought possible. That is one reason I am not to interested in it.

You have a great opportunity to get your foot in the door as a trusted entity. Once the dozens of Taiwanese out fits go to town they would be looked at secondly to you or any product your involved with.

I think I am really at a point testing water blocks is all I have to offer and I am not sure that would do much good with pH already doing a better job at it.

[Cathar]

I'm now having a little trouble comprehending the reason for the original question then.

You asked "Is it a lost art?"

I answered "No - there's still plenty to do", even for DIY'ers.

...and you respond saying that you're not interested.

I humbly put it to you that the "art" is not lost, but rather your interest in it is.

(Really tried hard to express that it a way that is direct to the point, but not meaning to offend directly - rather just as a point of introspection for you to consider).

[Edit: More independent testing is ALWAYS good though, IMO. Right now it's little more than a one-and-a-half horse show. If that's all that you are willing to offer, then that already alone is priceless]

[jaydee]

Originally Posted by Cathar

I'm now having a little trouble comprehending the reason for the original question then.

You asked "Is it a lost art?"

I answered "No - there's still plenty to do", even for DIY'ers.

...and you respond saying that you're not interested.

I humbly put it to you that the "art" is not lost, but rather your interest in it is.

(Really tried hard to express that it a way that is direct to the point, but not meaning to offend directly - rather just as a point of introspection for you to consider).

[Edit: More independent testing is ALWAYS good though, IMO. Right now it's little more than a one-and-a-half horse show. If that's all that you are willing to offer, then that already alone is priceless]

What I am saying is there are many less people interested in doing the work to make something custom. Which in turn makes it a lost art if hardly anyone is doing it anymore.

I am not arguing there isn't stuff the DIY CAN do, just saying there are less people that want to be a DIY that WILL do.

What I ment by not interested is I can't build a good water pump, I can't build a good radiator, I already built the best block my tooling will allow, and I have no reason to make a chiller (especially being I don't even water cool a PC anymore). I am not denying I lost interest, just seems to me a lot of the DIY crowd has aswell as there just isn't enough reason to be a DIY'er anymore. The real "art" came from the DIY crowd IMO.

Not sure if that made sense.

[Cathar]

Originally Posted by jaydee116

What I am saying is there are many less people interested in doing the work to make something custom. Which in turn makes it a lost art if hardly anyone is doing it anymore.

...and I would say that this is a consequence of the early adopter phenomenon.

Whenever something new (to explore) comes along, there's always those who will leap in boots'n'all and explore it while others sit back at home and say that it's all too scary.

These same pioneers are the ones responsible for developing the art. The art is not lost, it's just got commercial. The "art" lives on in every one of those pioneers, but many of them have moved on and the "torch" is passed on to those who will drive it to the next level.

I guess what I'm saying is that the "art" is not lost, but the "pioneering" stage is over, and maybe it is really this that you lament the loss of.

All technologies go through the stages of discovery/pioneering, transition to mass acceptance/production, then ongoing sustaining and ubiquity. The reason why there's less pioneer's about is because there's less and less to pioneer. The "art" is not gone, but rather the mystery, and it is the mystery that gets the pioneer's (what you're calling DIY'ers) involved.

It's somewhat related to the life-cycle of a company. The most exciting (and dangerous) times are always to be had at the beginning, and this attracts and requires a certain kind of people. Once that stage is over, these people get bored pretty quickly and move on, and leave the show to the mass producers/sustainers.

The show and the art goes on, just that the pioneers have left the stage.

[jaydee]

I won't argue with that. If I had $12,000 to spend on a new CNC mill or felt I could make money and invested in the $30,000 mill I want I would certainly do it and continue making new blocks. If I knew I could make money long term I would jump all over it. Problem is I don't see any money in it long term. Especially once the Orientals get in the game. I just can't justify spending more money on better tools to make blocks. Hard enough to justify spending the cash on the test bench. In all honesty I think I like testing blocks better than building them.

[maxSaleen]

I would say that waterblock design is close to a wall. Blocks like the MCW6000 and the WW are more than adequate in design and will be for years to come. Perhaps this is the wall for you, JD. From what I've seen of you on the forums, your specialty is in waterblocks. There isn't much left to do here....... not on CPUs anyways.
There are still plenty of blocks left to make. Mofsets, hard drives, PSU block. We all know it can be done. The challenge is to do so in a practical fashion.

Here is how it goes: an idividual, or DIYer, looks at a component to cool, firgures out how to do it, and publishes how to on the internet. Some other DIYers replicate it, report results, and eventually a manufacturer picks up on it. They mass produce products that benefit the rest of the community.

I don't think the goal of watercooling is necessarily in obtaining higher clock speeds, not anymore. Instead I think that watercooling's goal is to achieve silence. In this respect, much research needs to be conducted in passive rads.

[maxSaleen]

How DO the asians do it? I guess it's easy when you pay your employees $1 a day.

JD: have you done any work with 45degree inlets and outlets? According to some of the data pH dug up they should offer substantially less flow restriction than the 90degree inlets/outlets every commercial block uses at present.

[DryFire]

I don't think the art is lost. We still try to shove larger rads an pumps into smaller cases.

[Generic]

Agree 100% w/ maxSaleen.

I'm just a no-name lurker, but imho cooling all of a machine's components w/ water and doing it quietly with minimal to no fans is/should be the ultimate goal.

Unfortunately, that doesn't seem feasible until the indistry standardizes the mobo enough to eliminate the need for custom MOFSET blocks, etc.

Leave behind the CPU blocks and give us the ultimate case/radiator design!!

[jaydee]

Originally Posted by maxSaleen

How DO the asians do it? I guess it's easy when you pay your employees $1 a day.

JD: have you done any work with 45degree inlets and outlets? According to some of the data pH dug up they should offer substantially less flow restriction than the 90degree inlets/outlets every commercial block uses at present.

I think a center inlet with some type of jet is probably the best solution. In order for a center inlet with the jet to work is has to be strait down and not angled. I guess you could try and 45 the outlet somehow. You might gain some flow rate.

[jaydee]

Originally Posted by maxSaleen

How DO the asians do it? I guess it's easy when you pay your employees $1 a day.

Also an answer to this is shit like this: http://www.theinquirer.net/?article=22309

As I said they have stuff in R&D that we never thought possible just waiting for a jump in the market.

[pHaestus]

oh so I'm a half a horse now? As long as it's the back half then I agree

[Cathar]

Originally Posted by pHaestus

oh so I'm a half a horse now? As long as it's the back half then I agree

No, you were the full horse. OC.com was the other half. Am unsure which half, but probably the front, as in at least they're testing and one should never look a gift horse in the mouth...

[Jag]

With respect to the DIYers, there has been an evolution also.
Nowadays, you read that someone has access to a EDM wire machine or a CNC, things that in the past weren't common in terms of custom blocks.
In my opinion even the modding aspect is getting sort of "professional", because people want their ideas come to life but looking as good as a commercial product.

[pHaestus]

My reviews are incomplete and I'm sporadic (being kind here) about posting any content at all. Half a horse seems about right to me...

It occurs to me that you forgot about Systemcooling and coolingmasters. Both those sites publish pretty thorough wcing content, though the french can make digesting the articles at cooling masters a bit of a bore.

We should all merge into one uber-site; doubt anyone wants to give up control though.

[El]v[0IsEv1L]

Yes Chilled water/Pelt combo really helps...

Idle -34C to -40C
Load -20C to -25C

[jaydee]

Makes the file download faster?

[maxSaleen]

pH: I have talking to Diceman over at VH. They are trying to transform their site. You might talk to him about content/review sharing.

[El]v[0IsEv1L]

Makes the file download faster?

Lol....

No not really.. but it helps keep the temps down...

It's fun having an Athlon64 4250+ lol... It can prolly go much higher.. haven't upped the vcore yet.

BTW my Pentium4 3.2 EE on the same setup could only reach 4GHz. And it ran much warmer.

The CPU cooling specs are:

Swiftech MCW5002-64T™ (226W Pelt)
Eheim 1250 Pump
Criticool Waterplant res
Black Ice Xtreme II (mounted on top of case)
2x Panaflo High 120MM on a Vantec Rheobus on low (silent).


The VPU cooling specs are:

Swiftech MCW50-T (80W Pelt)
Maxi-Jet 1000 (265GPH with a built in res)
Black Ice Xtreme rev 2 (mounted on rear of case)
1x Panaflo High 120MM on a Vantec Rheobus on low (silent)


The Water Chiller specs are:

1x DangerDen Maze3-T (156W Pelt)
Via Aqua 1300 Pump
Black Ice Pro (mounted on front of case)
1x Panaflo High 120MM on a Vantec Rheobus on low (silent)



And it all fits inside an Antec P160 case. I'm using a Meanwell dedicated PSU for the pelts which sits on top of the case.

P.S. I went Swiftech, for the most part, because of the quality. I have found that they make some of the best Thermoelectric coolers out there. The Maze blocks are not as good it seems in removing heat off a larger surface area compared to the newer Swiftech blocks.

Also the support is awesome.

[jaydee]

Originally Posted by El]v[0IsEv1L

Lol....

No not really.. but it helps keep the temps down...

It's fun having an Athlon64 4250+ lol... It can prolly go much higher.. haven't upped the vcore yet.

BTW my Pentium4 3.2 EE on the same setup could only reach 4GHz. And it ran much warmer.

The CPU cooling specs are:

Swiftech MCW5002-64T™ (226W Pelt)
Eheim 1250 Pump
Criticool Waterplant res
Black Ice Xtreme II (mounted on top of case)
2x Panaflo High 120MM on a Vantec Rheobus on low (silent).


The VPU cooling specs are:

Swiftech MCW50-T (80W Pelt)
Maxi-Jet 1000 (265GPH with a built in res)
Black Ice Xtreme rev 2 (mounted on rear of case)
1x Panaflo High 120MM on a Vantec Rheobus on low (silent)


The Water Chiller specs are:

1x DangerDen Maze3-T (156W Pelt)
Via Aqua 1300 Pump
Black Ice Pro (mounted on front of case)
1x Panaflo High 120MM on a Vantec Rheobus on low (silent)



And it all fits inside an Antec P160 case. I'm using a Meanwell dedicated PSU for the pelts which sits on top of the case.

P.S. I went Swiftech, for the most part, because of the quality. I have found that they make some of the best Thermoelectric coolers out there. The Maze blocks are not as good it seems in removing heat off a larger surface area compared to the newer Swiftech blocks.

Also the support is awesome.

Sounds pretty kick ass! Good work.

[Pajax]

What about power consumption? Think about environment!

Most people have no use of cold computer. I thing WC will not spread among common users quickly. A-64 (for example) do not produce so much heat as A-XP...

Home made watercooling is much cheaper than commercial. It cost me 20$ (pump included) . Its performance is good (bit better than dd RBX) and I have spent great time developing and manufacturing.
OT: What is the best block? LR Storm G5?